basquetWi + New ticket
venus VENUS-207

Suppress dormant announces entirely (Elazar 1a)

Done normal plpm-llmmsgsrv-cc

Elazar ruled 1a 2026-07-12: kill successionNotify dormant-announce emit (hub.mjs:2111) to pm-llmmsgsrv-cc. Zero-actionable noise (34 total/28 in 7d). Suppress ANNOUNCE only — leave succession/auto-resume mechanism + any god-view/KPI reads intact. coder-llmmsgsrv building; PM bounces after audit-clear. Ref mem:hub-noise-reduction-initiative.

Sub-tickets

No sub-tickets.
+ Add sub-ticket

Questions

No questions.

Activity

  • wi-cli-venus created · 5w ago
  • wi-cli-venus commented · 5w ago
    coder head cfc3dc4 (standalone), audit requested. Discriminator confirmed multiplexed: successionNotify has 3 branches — (1) DORMANT liveMembers==0 [CUT], (2) PARKED-SEAT none-electable [left; never fired, ever], (3) REAL SUCCESSION newPm [left]. Cutting wholesale would've killed live PM handoffs. Only branch-1 INSERT deleted; RETURN + seat-preservation + auto-resume intact; zero downstream readers of the announce. Data: 24 dormant emits not 34 (all in 7d, 13 rooms). Gate=audit-clear of cfc3dc4 → PM bounce.
  • wi-cli-venus commented · 5w ago
    HEAD moved cfc3dc4→43d0ea1: audit found MED (stale contract doc docs/aro-topology-v3.md promised a meta-PM note the cut no longer sends). Fixed: doc splits dormant(silent)/parked-seat(announce) in a table; hub.mjs:3066 comment corrected; added trap-warning that successionNotify must be called UNCONDITIONALLY (dormant branch does the PARKING, not just the announce — gating skips seat-preservation). Parked-seat branch RESOLVED = reachable (PM-prune with spectator-only members → liveMembers>0 but pickPmCandidate filters spectators → newPm=null → branch 2); not dead code, stays. Mechanism independently audit-verified. Gate=audit-clear of 43d0ea1 → PM bounce.
  • wi-cli-venus commented · 5w ago
    HEAD 43d0ea1→725a5ea: coder self-retracted a FALSE trap-warning (had claimed 'dormant branch PARKS, gating skips seat-preservation'). Truth verified in source: electPm(aro){winner=pickPmCandidate(aro); if(winner) stmtAroSetPm.run(...); return winner} — null candidate = NO write, so aro_config.pm_agent stays sticky BEFORE successionNotify runs. successionNotify on null path is a PURE notifier, parks nothing. Seat preservation lives in pickPmCandidate/electPm null path. Call stays unconditional for the parked-seat NOTE, not for preservation. Fixed in both hub comment AND topology doc + explicit 'earlier version claimed dormant parks — it does not' anti-reintroduction note. coder asked audit to verify electPm null path is SOLE seat-preservation writer (resolvePm/join/prune don't also write/clear pm_agent). Gate=audit-clear of 725a5ea, NOT 43d0ea1.
  • wi-cli-venus completed · 5w ago
    Deployed + verified 2026-07-12 (hub v2.34.7, head 725a5ea, 3 commits cfc3dc4+43d0ea1+725a5ea on 8f8bafd). Dormant announce (branch-1 emit only) suppressed; branches 2/3 (parked-seat note, real succession) untouched. Verify: 0 new 'dormant — all members offline' rows since bounce (total frozen 33); orphan_join=0. succession + seat-resume mechanism audit-verified (electPm(null) writes nothing → aro_config.pm_agent stays sticky before successionNotify runs; successionNotify null-path is pure notifier). Contract doc + hub comment corrected (audit found + coder self-retracted a false 'dormant branch parks' claim across 2 rounds). Ref mem:dormant-aro-sop (3b superseded), mem:hub-noise-reduction-initiative.
task
5w ago by wi-cli-venus
5w ago
2026-07-12 06:52